The notion of Symmetry is indeed closely related to the notion of physical law. However IMO, the questioning of a deeper understandin of physical law, also then implies the quest for a deeper understanding of symmetry.
The common way of thinking, to picture - in a given fixed state space - a perfect symmetry in the sense that the laws of physics are invariant with respect to that, and that this symmetry is broken down in phase transitions as the energy scale changes, to produce a set of smaller symmetries, is still tied to the "old scheme" that I think is inadequate.
Just to explain how I see it: If we like to instead use the symmetry language, evolving law, means evolving symmetries, and the radical view I advocate and which I beleivce is in line with Ungers vision is not just a simple "breaking of a perfect symmetry". The reason is that the notion of symmetry is a bit complex too, becuase if you require that "knowing the symmetries" means having information about something, one must ask how the process looks like whereby this symmetry is a result of an inference. This thing is typically ignored in the standard scheme, if one tries to explain broken symmetries from a perfect fixed master symmetry.
To establish a symmetry, ie. to establish that some action is invariant with respect to certain transformations or complexions, one first need to establish the distinguishability of these complexions. And this is actuqally qa bit paradoxal, because if you have a strict invariance, then how is it possible to make an inference about this invariance? Something with the inference is wrong here. The conclusion seems to be that symmetries by nature are always unstable and evolving, and that a symmetry needs to be challanged in order to make sense.
To take an example, consider a blind man tell you that "I am indifferent to wether your clothes are red or blue". At first that may seem sensible, but the problem is that how can a blind man, in the first place, acquire the NOTION of COLOURS? This is not consistent.
It would be different is this guy had a history, where he used to see, and thus has a prior notion of something, that he by now have concluded is redundant (a symmetry).
Edit: instead the proper behaviour would isntead be that a blind man IS invariant with respect to your colours, BUT he would never launch such a statement! Rather if you mention the word red or blue, he just would understand you.
(Of course I'm simplifying here, but I think the point is illustrated; that the notion of INFERRING an INVARIANC is subtle)
This analysis of how an observer inferes, and relates to a symmetry, really connects the set of possible symmetries to the microstructure and makeup of the observer itself. Ie. symmetries are necessarily also observer dependent, in the sense that I think we should only talk about _inferrable_ symmetries.
In the standard framework, we talk about symmetries as existing in mathematical worlds, that are not subject to scientific inference process. To me this is a serious flaw.
I agree that evolving law, can similary be phrased in terms of evolving symmetries. But just like entropy observer dependent, so are laws and so are symmetries. And I mean observer dependent in a more general sense that JUST energy scale dependent. Energy scale does set constraints or limits, of the complexity of law - this is why inferrable laws are destined to unification (all beeing one indisitinguishable interaction) as the energy scale increases (which from the inside view means the opposite! the compleixty of inside observers goes to zero - this is why encoding complex diverse law just isn't possible from an information theoretic inference persepctive)